AA is the first thing I turn off on 4k. Dont miss it at all.
maybe we will never see them. Why would an AIB even invest money into vega? Simply no business case to be made.
yes, it was a very strange comment.
Also, I have no experience with this, but I've read around here several times and it seems the consensus is that 4K more or less eliminates the need for AA, right? You're basically absorbing a massive performance hit for something like 1% or less in visual improvement?
High pixel density reduces the need for AA. 4k gives you that at normal monitor sizes, but once you get up to 40" or more, the pixels are just as big as smaller low-res monitors, so the need for AA comes back....it seems the consensus is that 4K more or less eliminates the need for AA, right? You're basically absorbing a massive performance hit for something like 1% or less in visual improvement?
yes, it was a very strange comment.
Also, I have no experience with this, but I've read around here several times and it seems the consensus is that 4K more or less eliminates the need for AA, right? You're basically absorbing a massive performance hit for something like 1% or less in visual improvement?
The differences were revealed by AMD stopping in the packaging of the reviewsamples that delivered it to techmedia as an additional Vega chip chip complete with hbm2 memory on interposer. The Vega chips are not simply stuck on a substrate, but are mounted on an interposer that allows a quick connection to the hbm2 memory on that same interposer. The interposer then sits on a substrate that is soldered on the PCB of the video card. That chip package we have on the card along with the review picture put. Mr Hardware.info , who has shared our property since the acquisition of the Press Group, also received such a package, but with another loose chip as a surprise.
Hmm... this could be a big issue.
What is worse is, the height isn't flush, so, the heatsink would be stuck on crooked, and might lead to cracking.
https://tweakers.net/nieuws/128409/...e-hoogte-en-afwerkkwaliteit-aan-partners.html
Hmm... this could be a big issue.
What is worse is, the height isn't flush, so, the heatsink would be stuck on crooked, and might lead to cracking.
https://tweakers.net/nieuws/128409/...e-hoogte-en-afwerkkwaliteit-aan-partners.html
True, And good thing intel iGPU also sucks in terms of efficiency (performance/area, performance/watt). Also leaks indicate 800mhz GPU clocks for RR and low clocks, low voltage Vega might actually be very efficient.
For big Vega with HBM2 they had to meet performance targets to price it accordingly. So in typical AMD fashion the cards run miles beyond optimal efficiency. undervolt and underclock might actually make them into good mining cards...
Those cards sell like pancakes no matter what lol
people are too afraid not to get one
that beta driver for mining will push speculation high on what will be the final hashrate after drivers goes out of beta + miner optimization over time haha
https://hothardware.com/news/amd-radeon-rx-vega-mining-block-chain-ethereum
Vega 64 can draw around 275W of power per card, you'd be looking at 2000W PSUs, or five GPUs with a 1500W PSU, or a pair of 1000W PSUs
What is this big issue? That AMD is filling the gaps to protect the interposer? Wow this is worse than CNN and Foxnews.
Fury X most definitely didn't have more than it needs. It's well known that if you've got golden samples of HBM, you're in for a large boost.Improved Memory Bandwidth efficiency? That's a bit of a stretch. It may have improved a bit, or it could be that Fiji just had more than necessary.
Either way it isn't very good vs the competition. Vega has ~ 480 GB/s of memory bandwidth and it struggles against the GTX 1080 which has about 350 GB/s.
Raven Ridge will be a very nice part with a CPU section that is competitive with Intel, and an APU better than Intels iGP, but that is about it. Don't expect Raven Ridge to challenge event he lowest end NVidia mobile dGPU.
Getting your hopes up on something big happening in the Vega part of the Raven Ridge APU, is just setting up for another Vega disappointment.
The issue is that AMD is NOT filling the gaps on all chips.
Here is a retail card.
here it is with paste:
Nano should be at or slightly above 56 perfomance.Really im very curious how VEGA Nano will perform, will it manage to be faster than Fiji ???
I never turn it on at 4k.AA is the first thing I turn off on 4k. Dont miss it at all.
Nano should be at or slightly above 56 perfomance.
It got a TGP of 150W while it 56 has a TGP of 165W. Full dies tend to be more efficient at the same power targets, so I don't think the 15W will pose a problem for the Nano against 56.
To be honest I think with proper undervolting and GPU underclocking V56 can probably match Polaris efficiency.
Apparently they are getting packages from 2 different suppliers. One is filling in the gaps and one isnt.Is it me or is there no underfill in the one with the HBM2 chips sitting lower? That's gotta be an issue for chip longevity when changing heatsinks or cleaning thermal paste off it.... I don't think the microbumps can take much if any mechanical stress vs common solder balls which are much bigger.
What the hell... how did that leave the assembly line?
It isn't just the filling, since the ones that are filled are also level to Vega.Apparently they are getting packages from 2 different suppliers. One is filling in the gaps and one isnt.
It isn't just the filling, since the ones that are filled are also level to Vega.
Interesting.I've read Vega 56 has no filling and Vega 64 does.
It's perhaps strange if you haven't experienced it but you would understand when you have. I don't have perfect vision and wear glasses when gaming but jaggies are very apparent at 40" 4K whether you approve of it or not. I was on 980ti SLI, went Titan XP SLI and settled after 6 months on1080ti as Arma 3 was slower in SLI than not and I spend 95% of my gaming there. I like my games to look good and keep thinking about getting another 1080ti as GTA5 especially looked a lot better Titan XP/1080ti SLI so another 1080ti would do the trick but having owned that for 6 months it's really not worth £700/$850 to me for 5 hours better GTA5 per month.yes, it was a very strange comment.
Also, I have no experience with this, but I've read around here several times and it seems the consensus is that 4K more or less eliminates the need for AA, right? You're basically absorbing a massive performance hit for something like 1% or less in visual improvement?